TM1 Forum

Hi,

I am having some problems with TM1 Feeders between cubes, see http://forums.olapforums.com/viewtopic.php?f=3&t=3241

Gregory mentioned about the resource requirement for feeder statement in the source cube, in my case Company (3 elements), Scenario (7 ele), Year (7 years), Period (12 ele), Salesman (50), Customer (2000), product (2000). I think this kind of size is quite common in manufacturing. But a feeder that involve all Products, Customers, Salesmen and some year and periods will easily require over 20G RAM , as per Gregory. Does it mean that TM1 may not be appropriate for Manufacturing, esp those TM1 models that will have dimensions with over 2000 elements?

The Skipcheck/Feeders system is all about minimalism. If you are ever feeding to a consolidation of all products (2000+) or to the total level of the cube then I reccomend you dont use rules at all. I think that if you ever need to do a feeder like that you should use TI to write static values.

Hello,

First of all TM1 can be a really a powerful tool for manufacturing:). There are huge number of models with bigger dimensions you mentioned in your sample and usually less than 2 or 3 GB (so a 32 bit TM1) is enough for those models... and those models have not one but many business cubes.

Skipcheck / feeders let you tell TM1 which cells should be calculated in a spare cube. It is rather a business issue than a technical one.
When you understand the background of your business rules you can write better feeders.
Without them TM1 has to calculate every cell and therefore calculates tons of zero and the result will be a very slow system.

You can read a description of spare data in the following article:
http://www.bi-verdict.com/fileadmin/Fre ... htm#Sparse data consequences

If your cube is not spare but really dense then you need lots of RAM and over a certain degree of density you probably do not need Skipcheck / feeder but time and patience.

Kind Regards,
Peter

feeders really shouldn't be used to try and feed high level consolidations in a cube like that. i inhertied a model similar to that but with much bigger dimensions (one 25k+) which would crash the server everytime it tried to start after the users had started to put some real data in to it. TM1 has various ways to generate data, sometimes you need immediate results from a data change so a rule works. Other times you may have complicated data / calcs to do but data is relatively static, so a TI can work better. It's all about understanding the business needs, how the customer interacts with the models and building a solution that possibly compromises in some areas but overall, does the job it was intended to do in an acceptable manner.

TM1 is a perfectly acceptable tool for working with manufacturing data. I know some people using it to track hardware downtime and efficiency. We use it for resource planning in a very complicated business, it may primarily be used in the Finance world but it is far more adaptable than that, the key is understanding the best way to go about it.

Hi DeeDee

TM1 is an excellent tool for manufacturing, supply chain, etc. or any complex business modelling scenario.

If you are creating a totally dense cube then there is clearly something wrong with your design. Have a careful read of Gregor's response to your other post http://forums.olapforums.com/viewtopic. ... 688#p14676
The answer is in getting the design approach correct.